Latent class regression analysis for describing cognitive developmental phenomena: An application to transitive reasoning

The aim of cognitive developmental research is to explain latent cognitive processes or structures by means of manifest variables such as age, cognitive behaviour, and environmental influences. In this paper the usefulness of the latent class regression model is discussed for studying cognitive developmental phenomena. Using this model, the relationships between latent and manifest variables can be explained by means of empirical data without the need of strong a priori assumptions made by a cognitive developmental theory. In the latent class regression model a number of classes are distinguished which may be characterized by particular cognitive behaviour. Environmental influences on cognitive behaviour may vary for different (developmental) classes. An application is given of the latent class regression model to transitive reasoning data. The results showed that a Five-Class model best fitted the data and that the latent classes differ with respect to age, strategy use (cognitive behaviour) and the influence of task characteristics (environmental influences) on the strategy use. The flexibility of the model in terms of mixed measurement levels and treatment of different cognitive variables offers a broad application to several cognitive developmental phenomena.